Two-sided caregiver marketplace with AI matching

Two-sided marketplaces fail at the matching step. Parents could not efficiently evaluate caregivers, and caregivers could not surface themselves to the families they actually suited. Browsing a long list is not the same as finding the right match, and low match quality erodes trust in the platform itself.

• Treated matching—not listing volume—as the core product problem.
• Defined what "suitable" meant on both sides before designing any ranking.
• Built trust signals into the flow, since the category depends on confidence rather than price.

• Two-sided profiles with structured, comparable attributes
• AI-assisted matching and recommendation between families and caregivers
• Discovery flows designed around fit rather than list position
• Trust and verification signals surfaced at the decision point

Families reach a shortlist of genuinely suitable caregivers instead of browsing an undifferentiated list, and caregivers are surfaced to the families they fit.